Comment 5 for bug 1557172

Hi Joseph

I'll see if I can nab some extra hardware to try it out. Keep in mind that I haven't found a way to reproduce this reliably yet, with our usual setup it can take a few weeks before it happens. Basically we have a cron job that periodically exports the state of our machines, which includes calling lsusb. Eventually this bug happens, all future calls to the job get stuck in D, and the machine eventually runs out of resources, requiring it to be rebooted.

I did browse kernel sources upstream and didn't notice anything that would appear to fix this issue, but I'm certainly no kernel-guru.